Influence of reactor systems on the morphology of Aspergillus awamori. Application of neural network and cluster analysis for characterization of fungal morphology

Abstract

The development of germinating spores and the morphology of Aspergillus awamori were investigated in shake flasks, stirred tank and airlift tower loop reactors under various cultivation conditions. The particular morphological features, evaluated by digital image processing, were classifed using an artifical neural network. The input layer consisted of the morphological features and the cultivation time and the output layer consisted of four object types: globular pellets, elongated pellets, clumps and filamentous mycelia. The significance of the particular morphological features and their combination was determined by cluster analysis. The effect of temperature, phosphate concentration, rotation speed and number of baffles on the fungal morphology in shake flasks, the effect of phosphate concentration and impeller speed in a stirred tank and the effect of phosphate concentration in airlift tower loop reactors were investigated and compared. The relationship of the fungal morphology to process performance was discussed.